Quantization forces Bd = Ω(k ln N) for perfect top-k retrieval realizability, plus a B* = O(ln ln N) threshold below which no d works under uniform scalar quantization.
super hub Mixed citations
The Faiss library
Mixed citation behavior. Most common role is background (48%).
102
Pith papers citing it
65
external citations · Pith
Background
48% of classified citations
abstract
Vector databases typically manage large collections of embedding vectors. Currently, AI applications are growing rapidly, and so is the number of embeddings that need to be stored and indexed. The Faiss library is dedicated to vector similarity search, a core functionality of vector databases. Faiss is a toolkit of indexing methods and related primitives used to search, cluster, compress and transform vectors. This paper describes the trade-off space of vector search and the design principles of Faiss in terms of structure, approach to optimization and interfacing. We benchmark key features of the library and discuss a few selected applications to highlight its broad applicability.
hub tools
citation-role summary
background 11
method 10
baseline 1
other 1
citation-polarity summary
claims ledger
- abstract Vector databases typically manage large collections of embedding vectors. Currently, AI applications are growing rapidly, and so is the number of embeddings that need to be stored and indexed. The Faiss library is dedicated to vector similarity search, a core functionality of vector databases. Faiss is a toolkit of indexing methods and related primitives used to search, cluster, compress and transform vectors. This paper describes the trade-off space of vector search and the design principles of Faiss in terms of structure, approach to optimization and interfacing. We benchmark key features of
authors
co-cited works
representative citing papers
CO-LMLM externalizes factual knowledge into a vector-indexed text database during pretraining, retrieving via hidden-state queries, achieving higher factuality and lower perplexity than prior LMLMs and standard LMs at 135M–360M scale.
A Certify-then-Rectify wrapper around HNSW uses conformal risk control plus EVT-estimated graph stretch to deliver average-case HNSW speed with high-probability exact k-NN recovery.
NNN decoding selects documents via non-negative elastic net reconstruction of the query embedding, with a theorem showing it strictly dominates dense retrieval on correlated corpora and experiments showing gains over inner-product baselines.
Flash-GMM provides a memory-efficient GPU kernel for scalable GMM soft clustering that delivers 20x speedup, supports 100x larger datasets, and yields up to 1.7x fewer distance computations in ANN search.
The paper presents a design pattern that stores sharded Vamana ANN indexes inside Iceberg snapshots using Puffin files, enabling vector similarity search in compute-disaggregated engines while inheriting Iceberg atomicity and lifecycle management.
LLM-powered linters with CoT and RAG detect quantum programming problems more accurately than rule-based LintQ on Qiskit code, with higher precision, recall, and F1 scores.
Contrastive privacy is a new corpus-contrast test for semantic privacy in AI-sanitized media that uses latent concept measures and requires no manual labeling.
Modern text encoders resist second-order collapse under mean pooling because token embeddings concentrate tightly within texts, and this resistance correlates with stronger downstream performance.
CORAL uses an agentic loop to adaptively refine retrieval corpora and queries in multilingual RAG based on evidence critique, yielding up to 3.58 percentage point accuracy gains on low-resource language cultural QA benchmarks.
AsmRAG detects malware at 96% F1 and attributes families at 95% F1 by retrieving functionally similar assembly code via LLM embeddings and density-weighted anchor selection, remaining robust to metamorphic obfuscation.
MCI approximates dense nearest neighbor graphs via maximal clique covers and progressive local densification to support fast arbitrary-filtered approximate nearest neighbor search with reduced space.
SG-RAG frames retrieval as subgraph matching to ensure LLMs meet every condition in factual queries and reports large gains over baselines on a new 120k-pair ERQA dataset.
Adaptive trie-guided decoding with document context and tunable penalties improves in-document query auto-completion, outperforming baselines and larger models like LLaMA-3 on seen queries.
HORIZON creates a cross-domain, long-horizon user modeling benchmark from Amazon Reviews that tests generalization across time, domains, and unseen users, exposing gaps in sequential and LLM-based recommendation models.
LLM-based dense retrievers generalize better when instruction-tuned but pay a specialization tax when optimized for reasoning; they resist typos and corpus poisoning better than encoder-only baselines yet remain vulnerable to semantic perturbations, with larger models and certain embedding geometry,
SGA-MCTS distills MCTS trajectories into de-lexicalized State-Goal-Action atoms for hybrid retrieval, enabling open-weight LLMs to match frontier model performance on complex planning without fine-tuning.
JZ-Tree introduces a flattened Morton plane-based tree hierarchy enabling collaborative dual-tree walks that deliver more than 10x faster exact k-NN search and FoF clustering on GPUs for N greater than 10 million particles, with multi-GPU scaling.
Releases TencentGR-1M and TencentGR-10M datasets with baselines for all-modality generative recommendation in advertising, including weighted evaluation for conversions.
Benchmark study shows DCO methods for vector similarity search are not reliable silver bullets due to high sensitivity to data properties and hardware, making them unsuitable for production deployment.
Spectral Tempering derives an adaptive scaling factor γ(k) from the embedding eigenspectrum via local SNR analysis and knee-point normalization to achieve near-optimal compression without training or validation.
RNSG approximates the range-aware relative neighborhood graph (RRNG) to enable high-performance range-filtered ANN queries with one compact index instead of many.
TARG uses uncertainty scores from a short no-context draft to gate retrieval in RAG, matching Always-RAG accuracy while cutting retrievals by 70-90% on QA benchmarks.
TSGP uses a pre-trained transformer as a semantic variation operator in genetic programming, generalizes across d-dimensional problems, and outperforms standard GP, SLIM_GSGP, Deep Symbolic Regression, and Denoising Autoencoder GP on 24 datasets while producing more compact solutions.
citing papers explorer
-
What Limits Does Quantization Place on Dense Top-$k$ Retrieval? A Theoretical Study
Quantization forces Bd = Ω(k ln N) for perfect top-k retrieval realizability, plus a B* = O(ln ln N) threshold below which no d works under uniform scalar quantization.
-
Co-LMLM: Continuous-Query Limited Memory Language Models
CO-LMLM externalizes factual knowledge into a vector-indexed text database during pretraining, retrieving via hidden-state queries, achieving higher factuality and lower perplexity than prior LMLMs and standard LMs at 135M–360M scale.
-
HNSW with Accuracy Guarantees Using Graph Spanners
A Certify-then-Rectify wrapper around HNSW uses conformal risk control plus EVT-estimated graph stretch to deliver average-case HNSW speed with high-probability exact k-NN recovery.
-
Non-negative Elastic Net Decoding for Information Retrieval
NNN decoding selects documents via non-negative elastic net reconstruction of the query embedding, with a theorem showing it strictly dominates dense retrieval on correlated corpora and experiments showing gains over inner-product baselines.
-
Flash-GMM: A Memory-Efficient Kernel for Scalable Soft Clustering
Flash-GMM provides a memory-efficient GPU kernel for scalable GMM soft clustering that delivers 20x speedup, supports 100x larger datasets, and yields up to 1.7x fewer distance computations in ANN search.
-
Puffin-Backed Vector Indexes: Attaching Approximate Nearest Neighbor Indexes to Apache Iceberg Snapshots for Compute-Disaggregated Query Engines
The paper presents a design pattern that stores sharded Vamana ANN indexes inside Iceberg snapshots using Puffin files, enabling vector similarity search in compute-disaggregated engines while inheriting Iceberg atomicity and lifecycle management.
-
Beyond Rules: LLM-Powered Linting for Quantum Programs
LLM-powered linters with CoT and RAG detect quantum programming problems more accurately than rule-based LintQ on Qiskit code, with higher precision, recall, and F1 scores.
-
Contrastive Privacy: A Semantic Approach to Measuring Privacy of AI-based Sanitization
Contrastive privacy is a new corpus-contrast test for semantic privacy in AI-sanitized media that uses latent concept measures and requires no manual labeling.
-
Why Mean Pooling Works: Quantifying Second-Order Collapse in Text Embeddings
Modern text encoders resist second-order collapse under mean pooling because token embeddings concentrate tightly within texts, and this resistance correlates with stronger downstream performance.
-
CORAL: Adaptive Retrieval Loop for Culturally-Aligned Multilingual RAG
CORAL uses an agentic loop to adaptively refine retrieval corpora and queries in multilingual RAG based on evidence critique, yielding up to 3.58 percentage point accuracy gains on low-resource language cultural QA benchmarks.
-
AsmRAG: LLM-Driven Malware Detection by Retrieving Functionally Similar Assembly Code
AsmRAG detects malware at 96% F1 and attributes families at 95% F1 by retrieving functionally similar assembly code via LLM embeddings and density-weighted anchor selection, remaining robust to metamorphic obfuscation.
-
MCI: A Maximal Clique Index for Efficient Arbitrary-Filtered Approximate Nearest Neighbor Search
MCI approximates dense nearest neighbor graphs via maximal clique covers and progressive local densification to support fast arbitrary-filtered approximate nearest neighbor search with reduced space.
-
Structure Guided Retrieval-Augmented Generation for Factual Queries
SG-RAG frames retrieval as subgraph matching to ensure LLMs meet every condition in factual queries and reports large gains over baselines on a new 120k-pair ERQA dataset.
-
DocQAC: Adaptive Trie-Guided Decoding for Effective In-Document Query Auto-Completion
Adaptive trie-guided decoding with document context and tunable penalties improves in-document query auto-completion, outperforming baselines and larger models like LLaMA-3 on seen queries.
-
HORIZON: A Benchmark for In-the-wild User Behaviour Modeling
HORIZON creates a cross-domain, long-horizon user modeling benchmark from Amazon Reviews that tests generalization across time, domains, and unseen users, exposing gaps in sequential and LLM-based recommendation models.
-
On the Robustness of LLM-Based Dense Retrievers: A Systematic Analysis of Generalizability and Stability
LLM-based dense retrievers generalize better when instruction-tuned but pay a specialization tax when optimized for reasoning; they resist typos and corpus poisoning better than encoder-only baselines yet remain vulnerable to semantic perturbations, with larger models and certain embedding geometry,
-
SGA-MCTS: Decoupling Planning from Execution via Training-Free Atomic Experience Retrieval
SGA-MCTS distills MCTS trajectories into de-lexicalized State-Goal-Action atoms for hybrid retrieval, enabling open-weight LLMs to match frontier model performance on complex planning without fine-tuning.
-
JZ-Tree: GPU friendly neighbour search and friends-of-friends with dual tree walks in JAX plus CUDA
JZ-Tree introduces a flattened Morton plane-based tree hierarchy enabling collaborative dual-tree walks that deliver more than 10x faster exact k-NN search and FoF clustering on GPUs for N greater than 10 million particles, with multi-GPU scaling.
-
Tencent Advertising Algorithm Challenge 2025: All-Modality Generative Recommendation
Releases TencentGR-1M and TencentGR-10M datasets with baselines for all-modality generative recommendation in advertising, including weighted evaluation for conversions.
-
Distance Comparison Operations Are Not Silver Bullets in Vector Similarity Search: A Benchmark Study on Their Merits and Limits
Benchmark study shows DCO methods for vector similarity search are not reliable silver bullets due to high sensitivity to data properties and hardware, making them unsuitable for production deployment.
-
Spectral Tempering for Embedding Compression in Dense Passage Retrieval
Spectral Tempering derives an adaptive scaling factor γ(k) from the embedding eigenspectrum via local SNR analysis and knee-point normalization to achieve near-optimal compression without training or validation.
-
RNSG: A Range-Aware Graph Index for Efficient Range-Filtered Approximate Nearest Neighbor Search
RNSG approximates the range-aware relative neighborhood graph (RRNG) to enable high-performance range-filtered ANN queries with one compact index instead of many.
-
Retrieval as a Decision: Training-Free Adaptive Gating for Efficient RAG
TARG uses uncertainty scores from a short no-context draft to gate retrieval in RAG, matching Always-RAG accuracy while cutting retrievals by 70-90% on QA benchmarks.
-
Transformer Semantic Genetic Programming for d-dimensional Symbolic Regression Problems
TSGP uses a pre-trained transformer as a semantic variation operator in genetic programming, generalizes across d-dimensional problems, and outperforms standard GP, SLIM_GSGP, Deep Symbolic Regression, and Denoising Autoencoder GP on 24 datasets while producing more compact solutions.
-
Audio Flamingo 3: Advancing Audio Intelligence with Fully Open Large Audio Language Models
Audio Flamingo 3 introduces an open large audio-language model achieving new state-of-the-art results on over 20 audio understanding and reasoning benchmarks using a unified encoder and curriculum training on open data.
-
Zero-shot Concept Bottleneck Models
Z-CBMs achieve zero-shot interpretable predictions by retrieving concepts from a million-vocabulary web bank via cross-modal search and regressing labels with sparse linear regression.
-
Retrieval-Augmented Personalization with Foundation Models for Wearable Stress Detection
Retrieval from out-of-domain foundation models enables personalization of a lightweight transformer for stress detection, yielding +3.92% accuracy and +4.76% F1 gains on WESAD without user labels.
-
CodeTeam: An LLM-Powered Multi-Agent Framework for Repository-Level Code Generation
CodeTeam is an LLM multi-agent system that improves SketchBLEU by 4.1/2.9 points and achieves top test pass rates (34.6% PE, 42.3% SFT) on repository-level code generation benchmarks via role-specialized planning and implementation stages.
-
PACMS: Submodular Context Selection as a Pluggable Engine for LLM Agents
PACMS uses submodular context selection as a pluggable engine to choose relevant context from conversation, memory, and tool outputs instead of recency truncation.
-
Ghost Vectors: Soft-Deleted Embeddings Remain Reconstructible in HNSW Vector Databases
Soft-deleted HNSW embeddings are reconstructible from raw storage via Vec2Text inversion (up to 100% on structured data), addressed by epoch key rotation that reduces PII recovery to 0%.
-
Dash2Sim: Closed-Loop Driving Simulation from in-the-wild Dashcam Videos
Dash2Sim recovers metric geo-referenced 4D scenes from in-the-wild monocular dashcam videos to enable the ROADWork4D benchmark, revealing that current closed-loop planners fail on work zone lane changes.
-
ANNS-AMP: Accelerating Approximate Nearest Neighbor Search via Adaptive Mixed-Precision Computing
ANNS-AMP adapts distance-computation precision to vector-space regions via a lightweight cluster-level predictor and a bit-serial accelerator, delivering 163.76x/10.57x/2.06x average speedups and 1100x/39.41x/6.66x energy reductions versus CPU/GPU/custom baselines with <2.7% accuracy loss.
-
NTILC: Neural Tool Invocation via Learned Compression
NTILC replaces in-context tool registry lookup with learned latent retrieval using a signature-aware composite loss, reducing context consumption by over 95% and latency by up to 74%.
-
M2S-AVSR: Modality-aware Multi-view Self-supervised Representation for Robust Audio-Visual Speech Recognition
M2S-AVSR introduces multi-view self-supervised visual encoding and modality-aware fusion for AVSR, releases the AISHELL8-RealScene dataset, and reports relative gains up to 29.4% on LRS3 under perturbations plus new SOTA on MISP2021.
-
Towards Unified and Data-Efficient Prognostics and Health Management with Tabular Foundation Models
Tabular foundation models applied to PHM via signal-to-table conversion achieve the best average ranks across prognostic and diagnostic tasks and remain competitive in low-data regimes.
-
Slipstream: Locality-Aware Graph Index Construction for Streaming Approximate Nearest Neighbor Search
Slipstream exploits continuity in vector streams to reduce insertion costs in graph ANNS indexes via prior-insertion candidates and an adaptive controller, delivering up to 30.8x higher throughput at >=0.95 recall@10 on five datasets.
-
NBQ: Next-Best-Question for Dynamic Profiling
NBQ is a plug-and-play framework for adaptive question selection in conversations to produce structured user profiles, with QuickMatch enabling scalable reciprocal matching through approximate vector search.
-
ChildEval: When large language models meet children's personalities
ChildEval is a new benchmark with 29K child personas (ages 3-6) for evaluating LLMs on explicit and implicit preference following across daily life categories.
-
Diversed Model Discovery via Structured Table Discovery
StructuredSemanticSearch uses table discovery operators and orientation-aware integration on model-card tables to improve evidence coverage and diversity in model recommendation queries over a semantic baseline.
-
TIGER-FG: Text-Guided Implicit Fine-Grained Grounding for E-commerce Retrieval
TIGER-FG proposes text-guided implicit fine-grained grounding with dual distillation to address modality and granularity asymmetries in image-to-multimodal e-commerce retrieval, reporting Recall@1 gains of 6.1 and 34.4 points on two new benchmarks.
-
PipeANN-Filter: An Efficient Filtered Vector Search System on SSD
PipeANN-Filter improves filtered vector search latency and throughput on SSD by exploring a superset of valid vectors identified via probabilistic filters and verifying attributes only after selecting top-k candidates.
-
HiDream-O1-Image: A Natively Unified Image Generative Foundation Model with Pixel-level Unified Transformer
A pixel-space Diffusion Transformer with Unified Transformer architecture unifies image generation, editing, and personalization in an end-to-end model that maps all inputs to a shared token space and scales from 8B to over 200B parameters.
-
CustomerSim: Benchmarking and Aligning Multimodal Language Models as Retail User Simulators
A retail user-simulator benchmark and GRPO training recipe claim improved persona adherence, but the paper's abstract and body disagree on core numbers.
-
Similar Pattern Annotation via Retrieval Knowledge for LLM-Based Test Code Fault Localization
SPARK improves LLM-based test code fault localization by retrieving similar past faults and selectively annotating suspicious lines in new failing tests.
-
GASim: A Graph-Accelerated Hybrid Framework for Social Simulation
GASim accelerates hybrid LLM-ABM social simulations via graph-optimized memory, graph message passing, and entropy-driven agent grouping, delivering 9.94x speedup and under 20% token use while aligning with real-world trends.
-
Kernel Affine Hull Machines as Compute-Efficient Encoders for Frozen Semantic Spaces
KAHM yields a compute-efficient query encoder that outperforms matched learned adapters in reconstructing a frozen Mixedbread embedding space on an Austrian-law retrieval task while delivering an 8.53x CPU speedup.
-
Robust Multimodal Recommendation via Graph Retrieval-Enhanced Modality Completion
GRE-MC retrieves relevant subgraphs and uses a graph transformer plus sparse codebook to complete missing modalities, outperforming prior methods on recommendation benchmarks.
-
Efficient Multivector Retrieval with Token-Aware Clustering and Hierarchical Indexing
TACHIOM speeds up multivector retrieval by up to 247x in clustering and 9.8x in retrieval on MS-MARCOv1 and LoTTE benchmarks using token-distribution-aware centroid allocation and a graph-plus-PQ index, with comparable effectiveness to prior systems.
-
TwinGate: Stateful Defense against Decompositional Jailbreaks in Untraceable Traffic via Asymmetric Contrastive Learning
TwinGate deploys a stateful dual-encoder system with asymmetric contrastive learning to detect decompositional jailbreaks in untraceable LLM traffic at high recall and low false-positive rate with negligible latency.
-
DenseStep2M: A Scalable, Training-Free Pipeline for Dense Instructional Video Annotation
A scalable training-free pipeline using video segmentation, filtering, and off-the-shelf multimodal models creates DenseStep2M, a dataset of 100K videos and 2M detailed instructional steps that improves dense captioning, step grounding, and cross-modal retrieval.